Shagun Jhaver
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Shagun Jhaver has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Alyvia Walters
Rutgers University, USA
3 shared publications - Alice Qian Zhang
Carnegie Mellon University
2 shared publications Amy X. Zhang
University of Washington, USA
2 shared publications- Koustuv Saha
Siebel School of Computing and Data Science, University of Illinois Urbana-Champaign, Urbana, Illinois, USA
2 shared publications - Aditya Vashistha
Cornell University
1 shared publication Amy Bruckman
Georgia Institute of Technology, Atlanta, Georgia
1 shared publication
Latest Publications
- Open Access
- 10 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Social Computing
Cleaning Up the Streets: Understanding Motivations, Folk Theories, and Concerns of Users Flagging Social Media Content
Social media platforms offer flagging, a technical feature that empowers users to report inappropriate posts or bad actors to reduce online harm. This feature, available now on all major social media platforms, is enacted through deceptively simple interfaces that disguise complex underlying interactions among users, algorithms, and moderators. Through interviewing 25 social media users with prior flagging experience, we examine end-users’ understanding of flagging procedures, explore the factors that motivate them to flag, and surface their cognitive and privacy concerns. We found that a lack of procedural transparency in flagging mechanisms creates gaps in users’ folk theories about how the mechanisms operate, yet users recognize the importance of platforms providing flagging options. Our analysis highlights how flags raise critical questions about distributing labor and responsibility for addressing online harm between platforms and users. We recommend innovations in the flagging design space that enhance user comprehension, ensure privacy, and reduce cognitive burdens.
- 10 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Social Computing
Open AccessCleaning Up the Streets: Understanding Motivations, Folk Theories, and Concerns of Users Flagging Social Media Content
Social media platforms offer flagging, a technical feature that empowers users to report inappropriate posts or bad actors to reduce online harm. This feature, available now on all major social media platforms, is enacted through deceptively simple interfaces that disguise complex underlying interactions among users, algorithms, and moderators. Through interviewing 25 social media users with prior flagging experience, we examine end-users’ understanding of flagging procedures, explore the factors that motivate them to flag, and surface their cognitive and privacy concerns. We found that a lack of procedural transparency in flagging mechanisms creates gaps in users’ folk theories about how the mechanisms operate, yet users recognize the importance of platforms providing flagging options. Our analysis highlights how flags raise critical questions about distributing labor and responsibility for addressing online harm between platforms and users. We recommend innovations in the flagging design space that enhance user comprehension, ensure privacy, and reduce cognitive burdens.
- Open Access
- 10 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 690-694
COMPASS: The 2nd International Workshop on Socio-technical Approaches to Content Moderation and Platform Governance
- 10 October 2026
- Published by Association for Computing Machinery (ACM)
- p. 690-694
Open AccessCOMPASS: The 2nd International Workshop on Socio-technical Approaches to Content Moderation and Platform Governance
- Open Access
- 23 September 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 10 (6) , 1-33
Should the Powerful Be Moderated? Examining Social Media Users’ Perspectives on Moderating Politicians’ Inappropriate Speech
Prior HCI research has investigated social media users’ perspectives on the efficacy and fairness of content moderation tools, as well as their concerns around the ways that powerful individuals may not be subject to the same “rules” as other users. However, we argue that there is a sub-group of powerful accounts that deserve specialized attention: those of politicians. Through a focus group study, we outline user perspectives on how—or if—politicians’ inappropriate online speech should be moderated, and what mental models are influencing these perspectives. We find that participants’ concerns about platforms’ capitalist functions, the changed relationship between politics and media, the outsized influence of politicians’ speech, and First Amendment considerations all play heavily into their conceptions of the risks and benefits of moderating these powerful accounts. We also examine how participants perceive the benefits and shortcomings of different platform-enacted and user-directed moderation solutions to counteract inappropriate politician speech. Building upon our findings, we offer suggestions on how scholars of law and platform governance can help build design and policy solutions that better address this issue. We also discuss how HCI research can contribute to advancing the debates around it.
- 23 September 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 10 (6) , 1-33
Open AccessShould the Powerful Be Moderated? Examining Social Media Users’ Perspectives on Moderating Politicians’ Inappropriate Speech
Prior HCI research has investigated social media users’ perspectives on the efficacy and fairness of content moderation tools, as well as their concerns around the ways that powerful individuals may not be subject to the same “rules” as other users. However, we argue that there is a sub-group of powerful accounts that deserve specialized attention: those of politicians. Through a focus group study, we outline user perspectives on how—or if—politicians’ inappropriate online speech should be moderated, and what mental models are influencing these perspectives. We find that participants’ concerns about platforms’ capitalist functions, the changed relationship between politics and media, the outsized influence of politicians’ speech, and First Amendment considerations all play heavily into their conceptions of the risks and benefits of moderating these powerful accounts. We also examine how participants perceive the benefits and shortcomings of different platform-enacted and user-directed moderation solutions to counteract inappropriate politician speech. Building upon our findings, we offer suggestions on how scholars of law and platform governance can help build design and policy solutions that better address this issue. We also discuss how HCI research can contribute to advancing the debates around it.
- Open Access
- 23 May 2026
- Published by Wiley in Journal of the Association for Information Science and Technology
- Vol. 77 (8) , 1100-1113
Interpreting algorithmic information cues: User sensemaking of search autocomplete moderation
Autocomplete is a search feature that algorithmically generates information cues for any keywords entered in the search bar. While this feature makes the search process more efficient, it also frequently produces biased, misleading, offensive, or otherwise inappropriate suggestions. To address this problem, commercial search systems like Google Search now moderate Autocomplete information cues. However, we know relatively little about how users perceive this moderation process. Conducting interviews with 20 users of web search systems, I examine user attitudes toward the ethical tradeoffs in enacting moderation, reliance on search systems for making regulation decisions, and users' own role in moderating autocomplete information. My findings show that users desire greater visibility into the autocomplete moderation process. They also see flags and personal moderation mechanisms as promising avenues for themselves to exert greater agency within contemporary information infrastructures. My analysis bridges the fields of content moderation and search engine critiques, and lays the groundwork for enacting fair, accountable, and transparent search moderation.
- 23 May 2026
- Published by Wiley in Journal of the Association for Information Science and Technology
- Vol. 77 (8) , 1100-1113
Open AccessInterpreting algorithmic information cues: User sensemaking of search autocomplete moderation
Autocomplete is a search feature that algorithmically generates information cues for any keywords entered in the search bar. While this feature makes the search process more efficient, it also frequently produces biased, misleading, offensive, or otherwise inappropriate suggestions. To address this problem, commercial search systems like Google Search now moderate Autocomplete information cues. However, we know relatively little about how users perceive this moderation process. Conducting interviews with 20 users of web search systems, I examine user attitudes toward the ethical tradeoffs in enacting moderation, reliance on search systems for making regulation decisions, and users' own role in moderating autocomplete information. My findings show that users desire greater visibility into the autocomplete moderation process. They also see flags and personal moderation mechanisms as promising avenues for themselves to exert greater agency within contemporary information infrastructures. My analysis bridges the fields of content moderation and search engine critiques, and lays the groundwork for enacting fair, accountable, and transparent search moderation.
- Open Access
- 22 March 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Social Computing
- Vol. 9 (1) , 1-40
Incorporating Procedural Fairness in Flag Submissions on Social Media Platforms
Flagging mechanisms on social media platforms allow users to report inappropriate posts/accounts for review by content moderators. These reports are pivotal to platforms’ efforts toward regulating norm violations. This article examines how platforms’ design choices in implementing flagging mechanisms influence flaggers’ perceptions of content moderation. We conducted a survey experiment asking US respondents ( N =2,936) to flag inappropriate posts using one of 54 randomly assigned flagging implementations. After flagging, participants rated their fairness perceptions of the flag submission process along the dimensions of consistency, transparency, and voice (agency). We found that participants perceived greater transparency when flagging interfaces included community guidelines and greater voice when they incorporated a text box for open-ended feedback. Our qualitative analysis of open-ended responses highlights user needs for improved accessibility, educational support for reporting, and protections against false flags. We offer design recommendations for building fairer flagging systems without exacerbating the cognitive burden of submitting flags.
- 22 March 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Social Computing
- Vol. 9 (1) , 1-40
Open AccessIncorporating Procedural Fairness in Flag Submissions on Social Media Platforms
Flagging mechanisms on social media platforms allow users to report inappropriate posts/accounts for review by content moderators. These reports are pivotal to platforms’ efforts toward regulating norm violations. This article examines how platforms’ design choices in implementing flagging mechanisms influence flaggers’ perceptions of content moderation. We conducted a survey experiment asking US respondents ( N =2,936) to flag inappropriate posts using one of 54 randomly assigned flagging implementations. After flagging, participants rated their fairness perceptions of the flag submission process along the dimensions of consistency, transparency, and voice (agency). We found that participants perceived greater transparency when flagging interfaces included community guidelines and greater voice when they incorporated a text box for open-ended feedback. Our qualitative analysis of open-ended responses highlights user needs for improved accessibility, educational support for reporting, and protections against false flags. We offer design recommendations for building fairer flagging systems without exacerbating the cognitive burden of submitting flags.
- 27 December 2025
- Published by SAGE Publications in New Media & Society
- Vol. 28 (11) , 6089-6111
Examining how search engine users understand the production of Autocomplete suggestions
Autocomplete is a popular search feature that automatically generates query suggestions for any keywords entered in the search bar. In this research, I examine regular end-users’ folk theories of how general-purpose search engines produce such suggestions. Drawing on interviews with 20 search engine users, I found that users conceptualize Autocomplete as an automated agent that is influenced by three main factors: (1) searcher’s personal search history and profile, (2) aggregate population-wide queries, and (3) commercial advertising. Users’ assumption of these influences raises for them critical concerns about privacy, transparency, information insularity, targeted data manipulation, and the reproduction of societal biases in Autocomplete’s outputs. My analysis also shows that users view explanations as a promising mechanism to enhance accountability in Autocomplete systems. I highlight the factors that shape users’ mental models of Autocomplete and discuss how their algorithmic imaginaries stabilize platforms’ revenue models.
- 27 December 2025
- Published by SAGE Publications in New Media & Society
- Vol. 28 (11) , 6089-6111
Examining how search engine users understand the production of Autocomplete suggestions
Autocomplete is a popular search feature that automatically generates query suggestions for any keywords entered in the search bar. In this research, I examine regular end-users’ folk theories of how general-purpose search engines produce such suggestions. Drawing on interviews with 20 search engine users, I found that users conceptualize Autocomplete as an automated agent that is influenced by three main factors: (1) searcher’s personal search history and profile, (2) aggregate population-wide queries, and (3) commercial advertising. Users’ assumption of these influences raises for them critical concerns about privacy, transparency, information insularity, targeted data manipulation, and the reproduction of societal biases in Autocomplete’s outputs. My analysis also shows that users view explanations as a promising mechanism to enhance accountability in Autocomplete systems. I highlight the factors that shape users’ mental models of Autocomplete and discuss how their algorithmic imaginaries stabilize platforms’ revenue models.
- 20 November 2025
- Published by SAGE Publications in New Media & Society
- Vol. 28 (10) , 5571-5595
Moral disengagement and content moderation attitudes: Examining how apathy to online harms may disguise racially conservative beliefs
Social media users’ preferences for various content moderation interventions have been widely studied, but the implicit beliefs that connect to these preferences are less understood. Using a nationally representative survey data set, we investigate how end-users’ attitudes toward moderating harmful speech online relate to their offline racial attitudes. We find that racially conservative beliefs are significantly positively related to participants indicating a distaste for concepts related to content moderation and cancel culture, suggesting that racial conservatism may be a crucial factor to consider in assessing these attitudes. We discuss our findings through the lens of moral disengagement theory, positing that supporting “freedom of expression” by way of disagreeing with content moderation and cancel culture may be a contemporary mechanism of morally disengaging with the harmful effects of racially insensitive speech.
- 20 November 2025
- Published by SAGE Publications in New Media & Society
- Vol. 28 (10) , 5571-5595
Moral disengagement and content moderation attitudes: Examining how apathy to online harms may disguise racially conservative beliefs
Social media users’ preferences for various content moderation interventions have been widely studied, but the implicit beliefs that connect to these preferences are less understood. Using a nationally representative survey data set, we investigate how end-users’ attitudes toward moderating harmful speech online relate to their offline racial attitudes. We find that racially conservative beliefs are significantly positively related to participants indicating a distaste for concepts related to content moderation and cancel culture, suggesting that racial conservatism may be a crucial factor to consider in assessing these attitudes. We discuss our findings through the lens of moral disengagement theory, positing that supporting “freedom of expression” by way of disagreeing with content moderation and cancel culture may be a contemporary mechanism of morally disengaging with the harmful effects of racially insensitive speech.
- Open Access
- 16 October 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 9 (7) , 1-35
Understanding the Prevalence of Caste: A Critical Discourse Analysis of Caste-based Marginalization on X
Despite decades of anti-caste efforts, sociocultural practices that marginalize lower-caste groups in India remain prevalent and have even proliferated with the use of social media. This paper examines how groups engaged in caste-based discrimination leverage platform affordances of the social media site X (formerly Twitter) to circulate and reinforce caste ideologies. Using a critical discourse analysis (CDA) approach, we examine the rhetorical and organizing strategies of 50 X profiles representing upper-caste collectives. We find that these profiles leverage platform affordances such as information control, bandwidth, visibility, searchability, and shareability to construct two main arguments: (1) that their upper caste culture deserves a superior status and (2) that they are the ''true'' victims of oppression in society. These profiles' digitally mediated discursive strategies contribute to the marginalization of lower castes by normalizing caste cultures, strengthening caste networks, reinforcing caste discrimination, and diminishing anti-caste measures. Our analysis builds upon previous HCI conceptualizations of online harms and safety to inform how to address caste-based marginalization. We offer theoretical and methodological suggestions for critical HCI research focused on studying the mechanisms of power along other social categories such as race and gender.
- 16 October 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 9 (7) , 1-35
Open AccessUnderstanding the Prevalence of Caste: A Critical Discourse Analysis of Caste-based Marginalization on X
Despite decades of anti-caste efforts, sociocultural practices that marginalize lower-caste groups in India remain prevalent and have even proliferated with the use of social media. This paper examines how groups engaged in caste-based discrimination leverage platform affordances of the social media site X (formerly Twitter) to circulate and reinforce caste ideologies. Using a critical discourse analysis (CDA) approach, we examine the rhetorical and organizing strategies of 50 X profiles representing upper-caste collectives. We find that these profiles leverage platform affordances such as information control, bandwidth, visibility, searchability, and shareability to construct two main arguments: (1) that their upper caste culture deserves a superior status and (2) that they are the ''true'' victims of oppression in society. These profiles' digitally mediated discursive strategies contribute to the marginalization of lower castes by normalizing caste cultures, strengthening caste networks, reinforcing caste discrimination, and diminishing anti-caste measures. Our analysis builds upon previous HCI conceptualizations of online harms and safety to inform how to address caste-based marginalization. We offer theoretical and methodological suggestions for critical HCI research focused on studying the mechanisms of power along other social categories such as race and gender.
- 26 June 2025
- Published by SAGE Publications in New Media & Society
- Vol. 28 (7) , 3436-3459
Examining racial stereotypes in YouTube autocomplete suggestions
Autocomplete is a popular search feature that predicts queries based on user input and guides users to a set of potentially relevant suggestions. In this study, we examine what YouTube autocompletes suggest to users seeking information about race on the platform. Specifically, we perform an algorithm output audit of autocomplete suggestions for input queries about four racial groups and examine the stereotypes they embody. Using critical discourse analysis, we identify five major sociocultural contexts in which racial information appears – Appearance , Ability , Culture , Social Equity , and Manner . We found that the participatory nature of YouTube produces a multifaceted representation of race-related content in its search outputs, characterized by enduring historical biases, aggregated discrimination, and interracial tensions, while simultaneously depicting minority resistance and aspirations of a post-racial society. We call for innovations in content moderation policy design and enforcement to address existing racial harms in YouTube search outputs.
- 26 June 2025
- Published by SAGE Publications in New Media & Society
- Vol. 28 (7) , 3436-3459
Examining racial stereotypes in YouTube autocomplete suggestions
Autocomplete is a popular search feature that predicts queries based on user input and guides users to a set of potentially relevant suggestions. In this study, we examine what YouTube autocompletes suggest to users seeking information about race on the platform. Specifically, we perform an algorithm output audit of autocomplete suggestions for input queries about four racial groups and examine the stereotypes they embody. Using critical discourse analysis, we identify five major sociocultural contexts in which racial information appears – Appearance , Ability , Culture , Social Equity , and Manner . We found that the participatory nature of YouTube produces a multifaceted representation of race-related content in its search outputs, characterized by enduring historical biases, aggregated discrimination, and interracial tensions, while simultaneously depicting minority resistance and aspirations of a post-racial society. We call for innovations in content moderation policy design and enforcement to address existing racial harms in YouTube search outputs.
- Open Access
- 2 May 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 9 (2) , 1-25
Deplatforming Norm-Violating Influencers on Social Media Reduces Overall Online Attention Toward Them
From politicians to podcast hosts, online platforms have systematically banned (''deplatformed'') influential users for breaking platform guidelines. Previous inquiries on the effectiveness of this intervention are inconclusive because 1) they consider only a few deplatforming events; 2) they consider only overt engagement traces (e.g., likes and posts) but not passive engagement (e.g., views); 3) they do not consider all the potential places influencers impacted by the deplatforming event might migrate to. We address these limitations in a longitudinal, quasi-experimental study of 165 deplatforming events targeting 101 influencers. We identify deplatforming events through Reddit posts and then manually curate the data, ensuring the correctness of a large dataset of deplatforming events. Then, we link these events to Google Trends and Wikipedia page views, platform-agnostic measures of online attention that capture the general public's interest in specific influencers. Through a difference-in-differences approach, we find that deplatforming reduces online attention toward influencers. After 12 months, we estimate that online attention toward deplatformed influencers is reduced by -63% (95% CI [-75%,-46%]) on Google and by -43% (95% CI [-57%,-24%]) on Wikipedia. Further, as we study over a hundred deplatforming events, we can analyze in which cases deplatforming is more or less impactful, revealing nuances about the intervention. Notably, we find that both permanent and temporary deplatforming reduces online attention toward influencers and that deplatforming influencers from multiple platforms further reduces the online attention they receive. Overall, this work contributes to the ongoing effort to map the effectiveness of content moderation interventions, driving platform governance away from speculation.
- 2 May 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 9 (2) , 1-25
Open AccessDeplatforming Norm-Violating Influencers on Social Media Reduces Overall Online Attention Toward Them
From politicians to podcast hosts, online platforms have systematically banned (''deplatformed'') influential users for breaking platform guidelines. Previous inquiries on the effectiveness of this intervention are inconclusive because 1) they consider only a few deplatforming events; 2) they consider only overt engagement traces (e.g., likes and posts) but not passive engagement (e.g., views); 3) they do not consider all the potential places influencers impacted by the deplatforming event might migrate to. We address these limitations in a longitudinal, quasi-experimental study of 165 deplatforming events targeting 101 influencers. We identify deplatforming events through Reddit posts and then manually curate the data, ensuring the correctness of a large dataset of deplatforming events. Then, we link these events to Google Trends and Wikipedia page views, platform-agnostic measures of online attention that capture the general public's interest in specific influencers. Through a difference-in-differences approach, we find that deplatforming reduces online attention toward influencers. After 12 months, we estimate that online attention toward deplatformed influencers is reduced by -63% (95% CI [-75%,-46%]) on Google and by -43% (95% CI [-57%,-24%]) on Wikipedia. Further, as we study over a hundred deplatforming events, we can analyze in which cases deplatforming is more or less impactful, revealing nuances about the intervention. Notably, we find that both permanent and temporary deplatforming reduces online attention toward influencers and that deplatforming influencers from multiple platforms further reduces the online attention they receive. Overall, this work contributes to the ongoing effort to map the effectiveness of content moderation interventions, driving platform governance away from speculation.